CN109389680A - Catchment basin of debris flow vital ground ratio characteristics screening technique - Google Patents
Catchment basin of debris flow vital ground ratio characteristics screening technique Download PDFInfo
- Publication number
- CN109389680A CN109389680A CN201811263728.8A CN201811263728A CN109389680A CN 109389680 A CN109389680 A CN 109389680A CN 201811263728 A CN201811263728 A CN 201811263728A CN 109389680 A CN109389680 A CN 109389680A
- Authority
- CN
- China
- Prior art keywords
- independent variable
- basin
- debris flow
- ratio characteristics
- catchment basin
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T17/00—Three dimensional [3D] modelling, e.g. data description of 3D objects
- G06T17/05—Geographic models
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A10/00—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE at coastal zones; at river basins
- Y02A10/40—Controlling or monitoring, e.g. of flood or hurricane; Forecasting, e.g. risk assessment or mapping
Landscapes
- Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Geometry (AREA)
- Software Systems (AREA)
- Remote Sensing (AREA)
- Computer Graphics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The defect of the comprehensive function effect of specific discriminant criterion and each factor cannot be provided for the prior art, the present invention discloses a kind of catchment basin of debris flow vital ground ratio characteristics screening technique.This method first acquires possible vital ground ratio characteristics collection from survey region catchment basin of debris flow sample and non-catchment basin of debris flow sample, then Lasso penalty is established, the optimal solution of penalty is solved, feature corresponding to the regression coefficient of β ≠ 0 is selected vital ground ratio characteristics in optimal solution feature.The present invention also provides the applications of catchment basin of debris flow vital ground ratio characteristics screening technique.The method of the present invention is the method based on Lasso method comparative analysis screening catchment basin of debris flow vital ground ratio characteristics, quantitative analysis is completed using machine learning algorithm and program, overcomes the drawbacks of prior art can only qualitatively screen vital ground ratio characteristics due to based on statistical method.Method And Principle is reliable, and as a result accuracy is high.
Description
Technical field
The present invention relates to a kind of terrain factor Feature Selection methods, are different from more particularly to a kind of screening catchment basin of debris flow
The method of the vital ground ratio characteristics of non-catchment basin of debris flow belongs to geology and geomorphology mapping exploratory techniques, hazards control skill
Art field.
Background technique
A kind of natural phenomena that mud-rock flow occurs as basin evolution to certain phase, matter and energy condition are to determine
It is developed and the fundamental factor of distribution, and matter and energy condition is decided by watershed unit condition to the full extent.Mud-rock flow one
As be small watershed event, thus the local feature in basin can generate conclusive influence, the cunning of many small watersheds to hydrologic process
The substance that slope, avalanche generate, defeated in the form of mud-rock flow to move on to outside basin often through cheuch, forming slope-paddy coupling is
System, and mud-rock flow is then a kind of performance of this coupled relation.All orographic conditions have area, have occur mud-rock flow can
Energy.The features of terrain of typical catchment basin of debris flow and non-catchment basin of debris flow is analyzed by comparing, and is removed uncorrelated features, be can determine
Determine the basin vital ground feature (hereinafter referred to as catchment basin of debris flow vital ground ratio characteristics) of Debris Flow Evolution.Mudstone is grasped
Basin vital ground ratio characteristics are flowed, the catchment basin of debris flow terrain study that building is more simply and readily explained on the one hand is conducive to
Model sentences knowledge for the catchment basin of debris flow based on formation mechenism and study on monitoring lays the foundation；It on the other hand can be to not yet occurring
Complete investigation is implemented in the basin for crossing mud-rock flow, and investigation distinguishes catchment basin of debris flow and non-catchment basin of debris flow.Two aspects are all to mud-rock flow
Diaster prevention and control is of great significance.
The prior art to the screening of catchment basin of debris flow vital ground ratio characteristics is divided using the means of mathematical statistics
Analysis research (Coe, J.A.and Godt, J.W., 2001, Debris flows triggered by the El
rainstorm of February 2-3,1998,Walpert Ridge and Vicinity；Zheng's Xu Zhi rainfall and ground text item
Research [D] TaiWan, China that part and earth flow are related: National Cheng Kung University's Master's thesis, 2003；Zeng Yichao earth flow occurs
Research [D] TaiWan, China of rainfall and the comprehensive Warning Index of ground text: National Cheng Kung University's Master's thesis, 2004；Liu,C.N.,
Huang H.F.,Dong J.J. Impacts of September 21,1999Chi-Chi earthquake on the
characteristics of gully-type debris flows in central Taiwan.Natural Hazards,
2008,47.).These researchs are restricted by technological means etc., currently more to the screening of catchment basin of debris flow vital ground ratio characteristics
It by virtue of experience determines, belongs to qualitative analysis.Each factor is individually divided using the method concentration of traditional Mathematical Statistics Analysis simultaneously
Analysis.The prior art could not provide the comprehensive function effect of specific discriminant criterion and each factor, it is difficult to which form unification has generation
The conclusion of table.
Summary of the invention
The purpose of the present invention is to the deficiencies in the prior art, provide a kind of based on Lasso method comparative analysis screening
The method of catchment basin of debris flow vital ground ratio characteristics, this method complete quantitative analysis using machine learning algorithm and program, really
Determine the vital ground ratio characteristics that catchment basin of debris flow is different from non-catchment basin of debris flow.
To achieve the above object, present invention firstly provides a kind of catchment basin of debris flow vital ground ratio characteristics screening technique,
Its technical solution is as follows:
A kind of catchment basin of debris flow vital ground ratio characteristics screening technique, for determining that catchment basin of debris flow is different from non-mudstone
Flow the vital ground ratio characteristics in basin, it is characterised in that:
Step S1, survey region delimited, investigation obtains O catchment basin of debris flow sample and P non-mudstones in survey region
Basin sample is flowed, O is equal or different to P, respectively two class sample labeling basin labels；
Step S2, investigation obtains the watershed unit characteristic values of two class samples respectively, record Q two class basin sample landform because
Subcharacter is the possibility vital ground ratio characteristics that catchment basin of debris flow is different from non-catchment basin of debris flow, and Q is not more than the sum of O and P, right
Each possibility vital ground ratio characteristics value constitutes possible vital ground ratio characteristics collection A after making centralized criterion processing；
Step S3, it according to Lasso criterion, establishes using vital ground ratio characteristics possible in set A as independent variable X, basin
Label is the penalty of dependent variable Y；
Step S4, the optimal solution that the penalty is solved with the minimum standard of penalty obtains penalty oneself
Variable active set B；
Step S5, feature corresponding to the regression coefficient of β ≠ 0 in independent variable active set B is put into set C, obtained crucially
Shape ratio characteristics collection C；Feature in set C is the vital ground factor spy that catchment basin of debris flow is different from non-catchment basin of debris flow
Sign.
In screening catchment basin of debris flow vital ground ratio characteristics, in face of the problem of mainly catchment basin of debris flow landform answer
There are complicated relationship between polygamy and mass data, it is easy to make to construct the morbid state that relational matrix becomes sensitive to disturbance
Battle array, not can be carried out effective Feature Selection.The above method of the present invention is realized using the Lasso method of machine learning.It thinks substantially
Think it is (in the sum of regression coefficient absolute value less than under the constraint condition of a constant, to make residual sum of squares (RSS) most according to Lasso rule
Smallization) one is constructed using possible vital ground ratio characteristics as independent variable, using basin attribute as the penalty of dependent variable, lead to
It crosses and trains, the model more refined is obtained under constraint condition.Through this process, it can preferably solve conllinear due to sample
Property be easy to make matrix to become the ill conditioned matrix sensitive to disturbance, in turn result in the unstable of regression coefficient analytic solutions, the parameter solved
Usually there is the problem of sparse feature, i.e., the corresponding parameter of many features can be zero.Make it possible feature selecting in this way,
The features of terrain most influenced on result can be filtered out.
In above method step S1, mud-rock flow in research area's digital elevation model (DEM) Research on partition region can use
Basin sample and non-catchment basin of debris flow sample, can also be around the substance and energy condition for forming mud-rock flow, by field investigation, distant
Feel image interpretation means and divides catchment basin of debris flow sample and non-catchment basin of debris flow sample.
In above method step S3, Lasso penalty is:
The formula of t >=03
Wherein, p-possibility vital ground ratio characteristics sum, t-given constrained parameters, N-catchment basin of debris flow sample
With catchment basin of debris flow total sample number, yiBasin label corresponding to i-th of sample of-dependent variable Y, xij- the i-th row jth column
Independent variable X value, β=(β1,…,βj,…,βp)TTreated that each possible vital ground ratio characteristics wait resolving for-centralized criterion
Regression coefficient, βjRegression coefficient to be resolved, β corresponding to-j-th possible vital ground ratio characteristics0- constant term.
In above method step S4, according to Lasso rule, initial related coefficient is first set as 0, is first found out and residual values
The maximum variable of related coefficient, standard are to keep penalty minimum, which is set as x1, add it to independent variable active set
B, then in x1A longest step-length is found in direction, so that there is next variable x2With the related coefficient and x of residual error1With it is residual
The related coefficient of difference is equal, by x2Active set is added.With same method, sequentially find in surplus variable with front addition activity
The variable of collection and the equal variable of the related coefficient of residual error, are added active set B.
The catchment basin of debris flow vital ground ratio characteristics filtered out using the above method are from the crisscross friendship of many possible features
It is determined in wrong correlation via quantitative analysis, overcomes tradition from quantity statistics method and independent analysis is carried out to each factor
Disadvantage, thus it is more accurate.The vital ground ratio characteristics that the method for the present invention filters out are committed to basin in survey region
Study on Mathematic Model also can provide basis to obtain the more accurate mathematic(al) structure of survey region landform.Therefore the present invention also provides
Application of the above-mentioned catchment basin of debris flow vital ground ratio characteristics screening technique in catchment basin of debris flow research.
Compared with prior art, the beneficial effects of the present invention are: (1) provides one kind based on the comparative analysis of Lasso method
The method for screening catchment basin of debris flow vital ground ratio characteristics, this method complete quantitative point using machine learning algorithm and program
Analysis, determines that catchment basin of debris flow is different from the vital ground ratio characteristics of non-catchment basin of debris flow, overcomes the prior art because to unite
The drawbacks of vital ground ratio characteristics can only qualitatively be screened based on meter method.(2) this method is by building one with can
Energy vital ground ratio characteristics are that independent variable obtains one and more refine using basin attribute as the Lasso penalty of dependent variable
Model, model compression part coefficient and setting section coefficient is zero, to preferably solve conllinear due to sample
Property be easy to make matrix to become the ill conditioned matrix sensitive to disturbance, in turn result in the unstable of regression coefficient analytic solutions, the parameter solved
Usually there is the problem of sparse features, i.e., the corresponding parameter of many features can be zero.It ensure that the validity of Feature Selection.(3)
Provide application of the catchment basin of debris flow vital ground ratio characteristics screening technique of the present invention in catchment basin of debris flow research.
Detailed description of the invention
Fig. 1 is Goats in Liangshan Prefecture typical catchment basin of debris flow and non-catchment basin of debris flow distribution map within the border.
Fig. 2 a is feature convergence rate figure.
Fig. 2 b is the partial enlarged view of Fig. 2 a.
Fig. 3 is feature selecting probability graph.
Specific embodiment
With reference to the accompanying drawing, the preferred embodiment of the present invention is further described.
Embodiment one
As shown in FIG. 1 to 3, catchment basin of debris flow is screened within the border in autonomous prefecture, the Liangshan of Sichuan Province Yi nationality, distributed over Yunnan, Sichuan and Guizhou with the method for the present invention
Vital ground ratio characteristics.
1, vital ground ratio characteristics are screened
Step S1, basin early period is investigated
Using digital elevation model (DEM), the domestic basin of Goats in Liangshan Prefecture is divided, wherein including known typical mud-rock flow stream
Domain and non-catchment basin of debris flow (Fig. 1 is Goats in Liangshan Prefecture typical catchment basin of debris flow and non-catchment basin of debris flow distribution map within the border).Obtain mudstone
Flow basin sample O=679, non-catchment basin of debris flow sample P=512.For catchment basin of debris flow sample labeling basin label (+1),
For non-catchment basin of debris flow sample labeling basin label (- 1).
Step S2, the watershed unit factor values of two class samples are extracted respectively
Various watershed unit ratio characteristics Q=14 of 1191, two class basin sample are further obtained using DEM, are denoted as
Possible vital ground ratio characteristics.Q (=14) meets calculating no more than O (=679) and the smaller of P (=512) between the two
It is required that.14 terrain factor features are respectively: dispersed elevation (1), gradient energy (2), groove gradient (3), elevation variance (4),
Relative relief (5), ruling grade (6), mean inclination (7), gradient variance (8), Melton ratio (9), drainage area
(10), basin perimeter (11), basin length (12), form factor of basin (13), the basin shape factor (14).It extracts possibly
When shape ratio characteristics, for the terrain factor feature only occurred in a kind of basin sample, it can be arranged in another basin sample
It is 0.
Centralized criterion processing is made to each possible vital ground ratio characteristics value, by each characteristic value be converted into 0~1 it
Between, so that data value primitive character is converted into nondimensional amount.Possibility vital ground ratio characteristics after standardization constitute set A.
Step S3, penalty is established according to Lasso criterion
According to Lasso criterion, establish using in set A may vital ground ratio characteristics as independent variable X, basin label be
The penalty of dependent variable Y.
The formula of t >=03
In formula, p-possibility vital ground ratio characteristics sum, step S2 are determined,
T-given constrained parameters, it is determining in 0~1 tentative calculation,
N-catchment basin of debris flow sample and non-catchment basin of debris flow total sample number, step S2 is determining,
yiBasin label corresponding to i-th of sample of-dependent variable Y, step S1 is determining,
xijThe independent variable X value of-the i-th row jth column, step S2 is determining,
β=(β1,…,βj,…,βp)TTreated that each possible vital ground ratio characteristics are to be resolved for-centralized criterion
Regression coefficient,
βjRegression coefficient to be resolved corresponding to-j-th possible vital ground ratio characteristics,
β0- constant term, initial value 0.
Step S4, the optimal solution of penalty is sought with the minimum standard of Lasso penalty, and variable active set B is added；
Step S41, β=(β is enabled1,…,βj,…,βp)TThe initial value of each element is 0, and the initial value of independent variable active set B is
It is empty；
Step S42, so that the minimum standard of penalty, determining and residual errorRelated coefficient is maximum from change
Measure the i-th row element (x of Xi=(xi1,…,xij,…,xip)T, by independent variable xiTerrain factor value be added independent variable activity
Collect B；Residual errorIn,For regression fit value, initial value 0；
Advance along variable in independent variable active set B, it is point-by-point to increase independent variable xiCorresponding factor betai, to reduce independent variable
xiWith the related coefficient of residual error, at this time regression fit valueResidual error is
Continue to calculate new residual errorWith the related coefficient between respective variable X, with factor betaiPoint-by-point increase,
Independent variable xiWith new residual errorRelated coefficient be gradually reduced, until occur a new independent variable xkMeet inner productThen along vector xiWith xkAngular bisector direction advance, it is accordingly point-by-point to increase
Big βiWith βk, corresponding residual error is
So circulation finds third independent variable, determines the first independent variable and the second independent variable angle in three independents variable
First angular bisector, the second angular bisector of the second independent variable and third independent variable angle form angle along two angular bisectors
Angular bisector direction move point by point, the coefficient of three independents variable is adjusted, so that three independents variable and current in moving process
The related coefficient of residual error reduces and equal always；
Step S43, through the corresponding regression coefficient β of independent variable X each in above-mentioned successively point-by-point adjustment independent variable active set B,
Reduce each independent variable with the related coefficient of current residue and equal always, residual error isWherein n is the number of arguments in independent variable active set B；
Step S44, determine whether the corresponding nonzero coefficient of each independent variable becomes 0 in independent variable active set B, if so, should
The independent variable that coefficient is 0 is rejected from independent variable active set B, and continues to execute S45；If it is not, continuing to execute S45；
Step S45, the phase relation of the outer independent variable X of independent variable active set B each row element and current residue is determined whether there is
Number is equal to the case where related coefficient of each independent variable and current residue in independent variable active set B if so, continuing to execute S46；It is no
Then, it returns and executes S43；
Step S46, independent variable active set B is added in the row element in the independent variable X outside independent variable active set B；
Step S47, whether in independent variable X have not processed row element, if so, returning to step S43 if determining；If
It is no, end operation.
Step S5, the factor for meeting default screening rule in independent variable active set B is put into vital ground ratio characteristics collection
C, default screening rule is: for solved in step S4 each possible vital ground ratio characteristics regression coefficient β to be resolved=
(β1,…,βj,…,βp)T, the corresponding feature of nonzero element is put into set C in β；Feature in set C is catchment basin of debris flow area
Not in the vital ground ratio characteristics of non-catchment basin of debris flow.
After the method for the present invention is handled, according to each watershed unit feature convergence rate, (Fig. 2 a is feature convergence rate figure, figure
2b is the partial enlarged view of Fig. 2 a) and key feature select probability (Fig. 3 is feature selecting probability graph), determine mud-rock flow
The vital ground feature in basin 4 is respectively: dispersed elevation (1), elevation variance (4), the basin shape factor (8), Melton
ratio(14)。
2, SVM and k-fold crosscheck
The mud-rock flow of selection and non-catchment basin of debris flow features of terrain data are allocated as two classes: one kind is by the method for the present invention
The feature of selection, another kind of unselected primitive character.
Two category features are uniformly divided into K group respectively, it is using svm classifier method that each landform is special to every category feature data
Sign subset data does an inspection set respectively, remaining K-1 group subset data obtains K model as training set, with this K
Performance indicator of the classification accuracy average of the final inspection set of model as classifier under this K-fold.Compare two kinds of classification
Under performance indicator, so that it is determined that the reliability of feature selecting.
The characteristic use SVM classifier after the primitive character to each basin and selection carries out k-folk crosscheck respectively,
By test, k value takes 7, as a result as follows:
It is input with 4 features after selecting, classification results accuracy rate is 57.9818%；With unselected 14
A primitive character is input, and classification results accuracy rate is 57.9405%.It is examined according to result above, it can be seen that two kinds of results
Difference is little, therefore 4 Jing Guo selection feature should be exactly vital ground feature possessed by catchment basin of debris flow, be respectively as follows:
Dispersed elevation (1), elevation variance (4), the basin shape factor (8), Melton ratio (14).
Claims (9)
1. catchment basin of debris flow vital ground ratio characteristics screening technique, for determining that catchment basin of debris flow is different from non-catchment basin of debris flow
Vital ground ratio characteristics, it is characterised in that:
Step S1, survey region delimited, investigation obtains O catchment basin of debris flow sample and P non-mud-rock flow streams in survey region
Domain sample, O are equal or different to P, respectively two class basin sample labeling basin labels；
Step S2, the watershed unit characteristic value of two class basin samples is extracted respectively, records Q terrain factor of two class basin samples
Feature is the possibility vital ground ratio characteristics that catchment basin of debris flow is different from non-catchment basin of debris flow, and Q is not more than the sum of O and P, to each
Possible vital ground ratio characteristics value constitutes possible vital ground ratio characteristics collection A after making centralized criterion processing；
Step S3, it according to Lasso criterion, establishes using vital ground ratio characteristics possible in set A as independent variable X, basin label
For the penalty of dependent variable Y；
Step S4, the optimal solution that the penalty is solved with the minimum standard of penalty, obtains the independent variable of penalty
Active set B；
Step S5, the feature that the regression coefficient of feature in independent variable active set B is not 0 is put into set C, obtain vital ground because
Subcharacter collection C；Feature in set C is the vital ground ratio characteristics that catchment basin of debris flow is different from non-catchment basin of debris flow.
2. according to the method described in claim 1, it is characterized by:
In the step S1, the basin label of catchment basin of debris flow sample labeling is+1, the basin label of non-catchment basin of debris flow sample
It is -1；
In the step S3, the penalty of building is:
The formula of t >=03
In formula, p-possibility vital ground ratio characteristics sum, step S2 are determined,
T-given constrained parameters, it is determining in 0~1 tentative calculation,
N-catchment basin of debris flow sample and non-catchment basin of debris flow total sample number, step S2 is determining,
yiBasin label corresponding to i-th of sample of-dependent variable Y, step S1 is determining,
xijThe independent variable X value of-the i-th row jth column, step S2 is determining,
β=(β1,…,βj,…,βp)TTreated each possible vital ground ratio characteristics recurrence to be resolved of-centralized criterion
Coefficient,
βjRegression coefficient to be resolved corresponding to-j-th possible vital ground ratio characteristics,
β0- constant term, initial value 0.
3. according to the method described in claim 2, it is characterized by: implementing in the step S4 according to following steps:
Step S41, β=(β is enabled1,…,βj,…,βp)TThe initial value of each element is 0, and the initial value of independent variable active set B is sky；
Step S42, so that the minimum standard of penalty, determining and residual errorThe maximum independent variable X's of related coefficient
I-th row element (xi=(xi1,…,xij,…,xip)T, by independent variable xiTerrain factor value be added independent variable active set B, institute
State residual errorIn,For regression fit value, initial value zero；
Advance along variable in independent variable active set B, it is point-by-point to increase independent variable xiCorresponding factor betai, to reduce independent variable xiWith
The related coefficient of residual error, at this time regression fit valueResidual error is
Continue to calculate new residual errorWith the related coefficient between respective variable X, with factor betaiPoint-by-point increase, from become
Measure xiWith new residual errorRelated coefficient be gradually reduced, until occur a new independent variable xkMeet inner productThen along vector xiWith xkAngular bisector direction advance, it is accordingly point-by-point to increase
Big βiWith βk, corresponding residual error is
So circulation finds third independent variable, determines first of the first independent variable and the second independent variable angle in three independents variable
Angular bisector, the second angular bisector of the second independent variable and third independent variable angle, along the angle of two angular bisector composition angles
Bisector direction is moved point by point, adjusts the coefficient of three independents variable, so that three independents variable and current residue in moving process
Related coefficient reduce and equal always；
Step S43, through the corresponding regression coefficient β of independent variable X each in above-mentioned successively point-by-point adjustment independent variable active set B, make certainly
Each independent variable reduces and equal always with the related coefficient of current residue in variable active set B, and residual error isWherein n is the number of arguments in independent variable active set B；Step S44, independent variable active set is determined
Whether the corresponding nonzero coefficient of each independent variable becomes 0 in B, if so, by the coefficient be 0 independent variable from independent variable active set B
It rejects, and continues to execute S45；If it is not, continuing to execute S45；
Step S45, each row element of the outer independent variable X of independent variable active set B and related coefficient of current residue etc. are determined whether there is
If so, continuing to execute S46 the case where the related coefficient of each independent variable and current residue in independent variable active set B；Otherwise, it returns
Receipt row S43；
Step S46, independent variable active set B is added in the row element in the independent variable X outside independent variable active set B；
Step S47, whether in independent variable X have not processed row element, if so, returning to step S43 if determining；If it is not,
End operation.
4. according to the method described in claim 3, it is characterized by: in the step S5, for solving each possibility in step S4
Vital ground ratio characteristics regression coefficient β=(β to be resolved1,…,βj,…,βp)T, the corresponding feature of nonzero element is in β
For vital ground ratio characteristics.
5. according to the method described in claim 1, it is characterized by: solving the optimal of the penalty in the step S4
Solution carries out constrained minimum to objective function using LASSO algorithm and calculates, and obtains regression vector.
6. according to the method described in claim 5, it is characterized by: being obtained by the modified LARS solution penalty
The independent variable active set B of penalty.
7. according to the method described in claim 1, it is characterized by: in the step S1, possible vital ground ratio characteristics
Including dispersed elevation, gradient energy, groove gradient, elevation variance, ruling grade, relative relief, form factor of basin, basin shape
The shape factor, gradient variance, mean inclination, basin perimeter, drainage area, basin length, Melton ratio.
8. according to the method described in claim 1, it is characterized by: being ground in the step S1 using digital elevation model division
Study carefully catchment basin of debris flow sample and non-catchment basin of debris flow sample in region, or around the substance and energy condition for forming mud-rock flow,
Catchment basin of debris flow and non-catchment basin of debris flow are divided by field investigation, remote sensing image interpretation means, obtains basin sample.
9. the claim 1~8 catchment basin of debris flow vital ground ratio characteristics screening technique is in catchment basin of debris flow research
Using.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811263728.8A CN109389680B (en) | 2018-10-28 | 2018-10-28 | Method for screening key topographic factor characteristics of debris flow basin |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811263728.8A CN109389680B (en) | 2018-10-28 | 2018-10-28 | Method for screening key topographic factor characteristics of debris flow basin |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109389680A true CN109389680A (en) | 2019-02-26 |
| CN109389680B CN109389680B (en) | 2022-06-07 |
Family
ID=65428146
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811263728.8A Active CN109389680B (en) | 2018-10-28 | 2018-10-28 | Method for screening key topographic factor characteristics of debris flow basin |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109389680B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110580573A (en) * | 2019-08-22 | 2019-12-17 | 广州地理研究所 | Debris flow risk assessment method and device based on terrain subsystem |
| CN110765418A (en) * | 2019-10-09 | 2020-02-07 | 清华大学 | Intelligent set evaluation method and system for basin water and sand research model |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102902870A (en) * | 2012-06-13 | 2013-01-30 | 中国科学院水利部成都山地灾害与环境研究所 | Debris flow forecasting method based on water and soil coupling in drainage basin |
| CN102999694A (en) * | 2012-10-30 | 2013-03-27 | 四川大学 | Debris flow risk evaluation method in area with frequent mountain disasters |
| CN103453944A (en) * | 2013-09-06 | 2013-12-18 | 中国科学院、水利部成都山地灾害与环境研究所 | Critical clear water flow measuring and calculating method for debris flow initiation and initiation monitoring method |
| CN104331744A (en) * | 2014-10-17 | 2015-02-04 | 中国科学院、水利部成都山地灾害与环境研究所 | Debris flow risk degree evaluation method |
| KR20160067660A (en) * | 2014-12-04 | 2016-06-14 | 부경대학교 산학협력단 | Vane type rheometer considering water flow characteristics, and method for obtaining rheological analysis technique using thereof |
| CN105975443A (en) * | 2016-05-04 | 2016-09-28 | 西南大学 | Lasso-based anomaly detection method and system |
| CN106223259A (en) * | 2016-09-28 | 2016-12-14 | 中国科学院、水利部成都山地灾害与环境研究所 | Mud-stone Flow of Gullies experimental system for simulating and the method determining Mud-stone Flow of Gullies start-up flow threshold value |
-
2018
- 2018-10-28 CN CN201811263728.8A patent/CN109389680B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102902870A (en) * | 2012-06-13 | 2013-01-30 | 中国科学院水利部成都山地灾害与环境研究所 | Debris flow forecasting method based on water and soil coupling in drainage basin |
| CN102999694A (en) * | 2012-10-30 | 2013-03-27 | 四川大学 | Debris flow risk evaluation method in area with frequent mountain disasters |
| CN103453944A (en) * | 2013-09-06 | 2013-12-18 | 中国科学院、水利部成都山地灾害与环境研究所 | Critical clear water flow measuring and calculating method for debris flow initiation and initiation monitoring method |
| CN104331744A (en) * | 2014-10-17 | 2015-02-04 | 中国科学院、水利部成都山地灾害与环境研究所 | Debris flow risk degree evaluation method |
| KR20160067660A (en) * | 2014-12-04 | 2016-06-14 | 부경대학교 산학협력단 | Vane type rheometer considering water flow characteristics, and method for obtaining rheological analysis technique using thereof |
| CN105975443A (en) * | 2016-05-04 | 2016-09-28 | 西南大学 | Lasso-based anomaly detection method and system |
| CN106223259A (en) * | 2016-09-28 | 2016-12-14 | 中国科学院、水利部成都山地灾害与环境研究所 | Mud-stone Flow of Gullies experimental system for simulating and the method determining Mud-stone Flow of Gullies start-up flow threshold value |
Non-Patent Citations (6)
| Title |
|---|
| DANIELA CASTROCAMILO ET AL.: "Handling high predictor dimensionality in slope-unit-based landslide susceptibility models through LASSO-penalized Generalized Linear Model", 《ENVIRONMENTAL MODELLING & SOFTWARE》 * |
| LUIGI LOMBARDO ET AL.: "Presenting logistic regression-based landslide susceptibility results", 《ENGINEERING GEOLOGY》 * |
| 张少杰等: "基于水文过程的泥石流预报中前期有效降水量的确定方法", 《水科学进展》 * |
| 张欢欢等: "基于PIE的高分遥感泥石流自动化变化检测方法研究", 《科技资讯》 * |
| 杨红娟等: "四川省泥石流灾害的时空分布规律和降水特征", 《灾害学》 * |
| 罗晓娟等: "基于数字高程模型的泥石流沟谷地形特征研究", 《西部探矿工程》 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110580573A (en) * | 2019-08-22 | 2019-12-17 | 广州地理研究所 | Debris flow risk assessment method and device based on terrain subsystem |
| CN110580573B (en) * | 2019-08-22 | 2021-04-30 | 广东省科学院广州地理研究所 | Debris flow risk assessment method and device based on terrain subsystem |
| CN110765418A (en) * | 2019-10-09 | 2020-02-07 | 清华大学 | Intelligent set evaluation method and system for basin water and sand research model |
| CN110765418B (en) * | 2019-10-09 | 2021-07-20 | 清华大学 | Intelligent set evaluation method and system for basin water and sand research model |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109389680B (en) | 2022-06-07 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Hermosilla et al. | Regional detection, characterization, and attribution of annual forest change from 1984 to 2012 using Landsat-derived time-series metrics | |
| Cipollina et al. | Reciprocal trade agreements in gravity models: A meta‐analysis | |
| Baudron et al. | Identifying the origin of groundwater samples in a multi-layer aquifer system with Random Forest classification | |
| Barua et al. | Artificial neural network–based drought forecasting using a nonlinear aggregated drought index | |
| Barua et al. | Comparative evaluation of drought indexes: case study on the Yarra River catchment in Australia | |
| Zilliox et al. | Tree species diversity and abundance as indicators of understory diversity in French mountain forests: Variations of the relationship in geographical and ecological space | |
| Arnaud et al. | Sensitivity of hydrological models to uncertainty in rainfall input | |
| Valbuena et al. | Characterizing forest structural types and shelterwood dynamics from Lorenz-based indicators predicted by airborne laser scanning | |
| Chen et al. | Monitoring soil organic carbon in alpine soils using in situ vis‐NIR spectroscopy and a multilayer perceptron | |
| Huang et al. | Predictive performance of ensemble hydroclimatic forecasts: Verification metrics, diagnostic plots and forecast attributes | |
| CN102930163A (en) | Method for judging 2 type diabetes mellitus risk state | |
| Wong et al. | High-resolution calculation of the urban vegetation fraction in the Pearl River Delta from the Sentinel-2 NDVI for urban climate model parameterization | |
| CN108595414A (en) | Heavy metal-polluted soil enterprise pollution source discrimination based on source remittance space variable reasoning | |
| CN105843829B (en) | A kind of big data creditability measurement method based on hierarchical mode | |
| WO2013009160A2 (en) | A geometric method for predicting landslide disaste | |
| Huang et al. | Local discrepancies in continental scale biomass maps: a case study over forested and non-forested landscapes in Maryland, USA | |
| CN106980750B (en) | A kind of Soil Nitrogen estimation method of reserve based on typical correspondence analysis | |
| Guo et al. | Variations in the precipitation–runoff relationship of the Weihe River Basin | |
| CN109389680A (en) | Catchment basin of debris flow vital ground ratio characteristics screening technique | |
| de Almeida Papa et al. | Evaluating tropical forest classification and field sampling stratification from lidar to reduce effort and enable landscape monitoring | |
| Hu et al. | Integrating CART algorithm and multi-source remote sensing data to estimate sub-pixel impervious surface coverage: a case study from Beijing Municipality, China | |
| Mustonen et al. | Defining the natural flow regimes of boreal rivers: relationship with benthic macroinvertebrate communities | |
| Jiang et al. | Mapping interannual variability of maize cover in a large irrigation district using a vegetation index–phenological index classifier | |
| Guisan et al. | Scaling the linkage between environmental niches and functional traits for improved spatial predictions of biological communities | |
| CN103499663B (en) | A kind of system of selection based on sensor in the Longjing tea Quality Detection Grade Model of genetic algorithm |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |